The Agent Skills Directory

[SAFE]: No prompt injection or safety bypass instructions were detected. The skill instructions are focused entirely on the legitimate task of stylistic editing and voice calibration.
[SAFE]: Data access is limited to the 'Read' tool for local reference files and user-provided voice profiles. There are no network tools or patterns suggesting data exfiltration or unauthorized credential harvesting.
[SAFE]: No obfuscation techniques such as Base64 encoding, zero-width characters, or homoglyph substitutions were found in the skill files.
[SAFE]: No external dependencies, remote script downloads, or dynamic code execution patterns (like eval or exec) were identified. The skill operates purely through text processing logic.
[SAFE]: Although the skill ingests external data from the user and local voice profiles, it lacks the high-privilege capabilities (such as shell execution or network access) that would be required to exploit an indirect prompt injection. This data ingestion is core to the skill's primary function and occurs within the agent's restricted environment.

humanize